SYNTHESIS AND PHOTOCATALYTIC ACTIVITY
OF THE MOS2 AND WS2 NANOPARTICLES
IN DEGRADATION OF ORGANIC COMPOUNDS
A THESIS
SUBMITTED TO THE GRADUATE SCHOOL
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE
MASTER OF SCIENCE
BY
DERAK JAMES
ADVISOR: TYKHON ZUBKOV
BALL STATE UNIVERSITY
MUNCIE, INDIANA
JULY, 2009
Abstract
Nanoparticles of MoS2 and WS2 were synthesized by decomposing the appropriate
metal hexacarbonyl in the presence of sulfur dissolved in decalin at 140°C. A significant
fraction of the nanoparticles was ≤ 15 nm in diameter as verified by Transmission
Electron Microscopy. The process was repeated in the presence of silica and then titania
to produce supported metal sulfides. The unsupported nanoparticles were found to
exhibit a size-dependent shift in their threshold UV-visible absorption due to quantum
confinement. Photocatalytic properties of each sulfide from synthesis in decalin were
explored by using each as a catalyst in the photodegradation of methylene blue by visible
light. These sulfides were also used to catalyze the photodegradation of acetone.
Unsupported MoS2 and WS2 nanoparticles catalyzed the photodegradation of acetone
under visible light of ≥ 400 nm wavelength. This is the first study reporting the
photocatalytic properties of the unsupported WS2 nanoparticles. Photodegradation of
methylene blue under ≥ 435 nm irradiation was detected using unsupported WS2 but not
unsupported MoS2, likely because activity was masked by the likely photobleaching of
the dye. When deposited on silica or titania, the nanosized MoS2 and WS2 could be
uniformly distributed in aqueous solutions to maximize the photocatalytic efficiency.
Correcting the absorbance measurements for light scattering by solids proved to be
beneficial for extracting kinetic information. Both silica deposited sulfides were found to
significantly increase the rate of methylene blue photodegradation, and deposited WS2
increased this rate significantly more than deposited MoS2. Similarly, both titania
deposited sulfides significantly increased the rate of methylene blue photodegradation,
and the deposited WS2 increased this rate significantly more than the deposited MoS2.
Table of Contents
Chapter 1. Introduction .......................................................................................................1
1.1. Semiconductor-Mediated Photocatalysis..................................................................1
1.2. Studies of the Photocatalytic Behavior of MoS2 and WS2 .......................................4
1.3. Carbonyl Based Synthesis of Nanosized MoS2 and WS2 .........................................7
1.4. Goals of the Reported Research................................................................................9
Chapter 2. Synthesis of the Sulfide Photocatalysts...........................................................10
2.1. Synthesis of Unsupported MoS2 and WS2 Nanoparticles in p-Xylene...................10
2.2. Decomposition of W(CO)6 in the Absence of Sulfur .............................................13
2.3. Synthesis of Unsupported MoS2 and WS2 Nanoparticles in Decalin .....................14
2.4. Synthesis of Silica Deposited MoS2 and WS2 Nanoparticles .................................16
2.5. Synthesis of Titania Deposited MoS2 and WS2 Nanoparticles...............................18
Chapter 3. Characterization of Synthesized Nanoparticles...............................................20
3.1. Imaging of Nanoparticles with Transmission Electron Microscopy ......................20
3.2. UV-visible Spectra of the Nanoparticles ................................................................28
3.3. Energy Dispersive X-Ray Analysis ........................................................................32
Chapter 4. Photocatalytic Degradation Tests: Setup and Protocols.................................36
4.1. Photocatalysis Setup ................................................................................................36
4.2. Photocatalytic Tests Protocols ................................................................................38
4.2.1. Methylene Blue Degradation by Unsupported MoS2 and WS2 ..........................39
4.2.2. Methylene Blue Degradation by MoS2 and WS2 Supported on SiO2 or TiO2Anatase................................................................................................................40
4.2.3. Accounting For Light Scattering by SiO2 and TiO2 ...........................................41
4.2.4. Acetone Methylene Blue Degradation By Unsupported MoS2 and WS2 ............45
Chapter 5. Photocatalytic Degradation Tests: Results.....................................................46
5.1. Methylene blue degradation by unsupported MoS2 and WS2 ..................................46
5.2. Methylene blue degradation by MoS2 and WS2 supported on SiO2 ........................49
5.3. Methylene blue degradation by MoS2 and WS2 supported on SiO2: Accounting
For Light Scattering by SiO2..................................................................................51
5.4. Methylene blue degradation by MoS2 and WS2 supported on TiO2-Anatase..........55
5.5. Methylene blue degradation by MoS2 and WS2 supported on TiO2-Anatase:
Accounting For Light Scattering by TiO2..............................................................56
5.6. Acetone Degradation By Unsupported MoS2 and WS2 ..........................................60
5.7. Conclusions..............................................................................................................63
References..........................................................................................................................65
i
List of Figures
Figure 1. Photocatalytic destabilization of an adsorbed molecule by TiO2........................1
Figure 2. Strategies for optimization of the bandgap for photocatalysis. ...........................3
Figure 3. Increase of the photocatalytic efficiency of composite semiconductors due to
improved charge separation across the interface. ..........................................................6
Figure 4. Dried products from the synthesis of unsupported MoS2 and WS2...................16
Figure 5. TEM Image of MoS2 in p-xylene at 60,000x magnification. ...........................21
Figure 6. TEM image of the Catalase standard at 200,000x magnification. ....................22
Figure 7. TEM image of MoS2 in p-xylene at 150,000x magnification. ..........................23
Figure 8. TEM image of WS2 in p-xylene at 100,000x magnification. ............................24
Figure 9. TEM image of MoS2 in decalin at 200,000x magnification..............................25
Figure 10. TEM image of WS2 in decalin at 200,000x magnification. ............................26
Figure 11. TEM image of WS2 in decalin at 200,000x magnification with high contrast
added. ...........................................................................................................................27
Figure 12. TEM image of WS2 in decalin at 100,000x magnification. ............................28
Figure 13. UV-visible difference spectra determined from spectra obtained during the
settling of MoS2 in p-xylene. .......................................................................................30
Figure 14. UV-visible difference spectra determined from spectra obtained during the
settling of WS2 in decalin. ...........................................................................................31
Figure 15. UV-visible difference spectra determined from spectra obtained during the
settling of MoS2 in decalin...........................................................................................32
Figure 16. Energy dispersive X-ray analysis of WS2 synthesized in decalin. ..................33
Figure 17. Energy dispersive X-ray analysis of MoS2 synthesized in decalin. ................34
Figure 18. Close-up view of the UV-visible light beam irradiating the beaker with a
reaction mixture. Long-pass cut-off filter is visible in the upper left corner..............37
Figure 19. The path of the light through the experiment setup.........................................37
Figure 20. (A) UV-visible spectrum of aqueous methylene blue. (B) UV-visible
spectrum of aqueous methylene blue in the presence of WS2/SiO2.............................42
Figure 21. An increase in scattering by the dispersing catalyst contributes an observable
increase in the 665 nm absorbance measurement as the mixture stirs over time. .......43
Figure 22. Relative UV-visible absorption at 665 nm plotted over time using
unsupported MoS2 and WS2 as the catalysts................................................................46
Figure 23. Methylene blue degradation by MoS2/SiO2, WS2/SiO2, and with the SiO2
control. .........................................................................................................................50
Figure 24. Linear correlation between light intensity losses at 665 nm and at 420 nm for
aqueous suspensions of three solids: (A) MoS2/SiO2, (B) WS2/SiO2, (C) SiO2. .......52
Figure 25. Methylene blue degradation by MoS2/SiO2, WS2/SiO2, and with the SiO2
control adjusted for light scattering. ............................................................................54
Figure 26. Relative UV-visible absorption at 665 nm plotted over time using titania and
titania deposited MoS2 and WS2 as catalysts...............................................................55
Figure 27. Linear correlation between light intensity losses due to scattering at 665 nm
and at 375 nm for aqueous suspensions of three solids: (A) MoS2/TiO2,
(B) WS2/TiO2, (C) TiO2..............................................................................................57
Figure 28. Methylene blue degradation by MoS2/TiO2, WS2/TiO2, and with the TiO2
control. . .......................................................................................................................59
ii
Figure 29. The 265 nm absorption peak in the baseline corrected spectrum of acetone
decreases as the acetone/MoS2 mixture is irradiated by light ≥ 400 nm......................61
Figure 30. Baseline corrected, relative UV-visible absorption at 265 nm plotted over
time using unsupported MoS2 and WS2 as the catalysts..............................................62
iii
List of Tables
Table 1. Reactant quantities for synthesis of unsupported metal sulfides in p-xylene. ....11
Table 2. Reactant quantities for control synthesis. ...........................................................13
Table 3. Reactant quantities for synthesis of unsupported metal sulfides in decalin. ......14
Table 4. Reactant quantities for synthesis of silica supported metal sulfides...................17
Table 5. Reactant quantities for synthesis of titania supported metal sulfides. ................18
Table 6. Quantities of solid used as catalysts to test deposited sulfides. ..........................40
Table 7. Quantities of solid used in tests to determine proportionate scattering. .............44
Table 8. Analysis of the rates of the methylene blue degradation by unsupported MoS2
and WS2. ......................................................................................................................48
Table 9. Analysis of the rates of the methylene blue degradation by MoS2/SiO2 and
WS2/SiO2......................................................................................................................51
Table 10. Functions relating the scatter at 665 nm and 420 nm for MoS2/SiO2, WS2/SiO2,
and SiO2. ......................................................................................................................51
Table 11. Analysis of the rates of the methylene blue degradation by MoS2/SiO2 and
WS2/SiO2......................................................................................................................54
Table 12. Analysis of the rates of the methylene blue degradation by MoS2/TiO2 and
WS2/TiO2. ....................................................................................................................56
Table 13. Functions relating the scatter at 665 nm and 375 nm for MoS2/TiO2, WS2/TiO2,
and TiO2. ......................................................................................................................58
Table 14. Error Analysis of the rates of the methylene blue degradation by MoS2/TiO2
and WS2/TiO2...............................................................................................................59
Table 15. Analysis of the rates of the acetone degradation by unsupported MoS2 and
WS2. .............................................................................................................................63
iv
Chapter 1. Introduction
1.1. Semiconductor-Mediated Photocatalysis
The predominant photocatalyst for pollutant degradation is currently TiO2.[1] TiO2,
or titania, is a wide-band semiconductor. When it absorbs light, an electron is excited
from the valence band to the conduction band[1] as illustrated by the left-hand diagram in
Figure 1.
Photocatalyst
Surface
Conduction Band
Adsorbed
Molecule
eEnergy
hν
Empty orbital
h+
Filled orbital
Valence Band
Figure 1. Photocatalytic destabilization of an adsorbed molecule by TiO2.
Reduction of
Adsorbed
Molecule
Oxidation of
Adsorbed
Molecule
The photogenerated electron and hole often recombine. However, they can also migrate
to the surface and appear in close proximity to an adsorbed species. The excited electron
can then reduce an adsorbed species by transferring to an available orbital. This available
orbital must be of lower energy compared to the energy of the conductance band. This is
a transfer of the electron from a more negative redox potential towards zero.[1] The
addition of an electron to the adsorbed species can convert the adsorbed molecule to
radicals or radical ions. These may undergo rearrangements, fragmentations, and
reactions with other molecules. Similarly, the adsorbed species can become oxidized by
transferring an electron from one of its occupied orbitals to the hole, h+, in the
semiconductor’s valence band. This is equivalent to the transfer of the hole to the
occupied orbital of the adsorbed molecule.[1] For the hole to transfer, the valence band
must be of lower energy than the occupied molecular orbital, i.e. the hole must have a
highly positive redox potential.
In water, TiO2 will usually act as a photocatalyst by oxidizing the water resulting in a
hydroxyl radical.[1] Water is oxidized when a hole in the valence band, h+, with
sufficient redox potential accepts an electron from the highest filled orbital. An H+ is lost,
and the resulting ●OH can diffuse in the water and react with other molecules. Organic
contaminants in the water will eventually come into contact with a hydroxyl radical,
resulting in oxidation. This process is capable of oxidizing a hydrocarbon to carbon
dioxide, water, and H+.[2, 3]
In order to absorb light for photocatalysis, an electron from the upper edge of the
valence band must absorb a photon with the energy of the bandgap. If the photon has
more energy than is required, then the excited electron will quickly relax through thermal
2
states to the lower edge of the conduction band. For titania, the required energy is about
3 eV, which is why titania is only capable of using about 3% of the sunlight reaching
Earth’s surface for photocatalysis .[4]
Two useful approaches to solve these problems are displayed in Figure 2. The first
one is to dope titania with other elements, metals or nonmetals. This can result in an
effective bandgap which is narrower than the original, thereby sensitizing titania to
visible light.
Strategy 1: Doping TiO2 to decrease the bandgap
Wide-Band Semiconductor
ConductanceBand
Conduction Ban
e-
“optimal bandgap”
Narrow-Band Semiconductor
ConductanceBand
Conduction Ban
Conduction Bad
ConductanceBand
e-
h+
h+
Valence Band
Energy
e-
h+
Valence Band
Valence Band
Strategy 2: Nanodispersing of narrow-band
semiconductors to widen the bandgap
Figure 2. Strategies for optimization of the bandgap for photocatalysis.
The second approach is to start with a completely different material, a narrow-band
semiconductor. Narrow-band semiconductors are capable of absorbing light throughout
the visible spectrum because they have much smaller bandgaps. However, these
bandgaps do not produce holes with high enough redox potentials to oxidize water.
Dispersing the narrow-band semiconductor on the nanoscale takes advantage of the
3
quantum size effect, resulting in a wider bandgap.[4,5] The quantum size effect arises in
particles of a size comparable to quantum dots.[6] Comparing these nanoparticles to the
bulk size, far fewer occupied and unoccupied orbitals contribute to the valence and
conductance bands respectively.[6] This leads to smaller bands with a wider bandgap
between them.
Exciting electrons across this wider bandgap results in a higher redox potential for the
hole. The hole may then be capable of oxidizing water to generate a hydroxyl radical.
The electron’s ability to absorb a photon with more than enough energy to cross the
bandgap results in a threshold frequency for absorbance. The increase in bandgap caused
by the quantum size effect has been demonstrated for MoS2.[4,5,7] It can be observed as
a blue-shift in the threshold absorption frequency of the nanoparticles compared to the
bulk metal sulfide. This blue-shift in the threshold absorbance of a sample of metal
sulfide indicates the presence of these nanoparticles.
1.2. Studies of the Photocatalytic Behavior of MoS2 and WS2
Studies of heterogeneous photochemistry on MoS2 have been pioneered by Tributsch
and Bennett.[8] However, the bulk of the well controlled experiments were done by
Wilcoxon et al.[4,5,7] Their work has shown that narrow-band semiconductors MoS2
and WS2 are capable of photodegradation of organic compounds using visible light. In
order to function as photocatalysts, the sulfides were either dispersed on the nanoscale or
used to sensitize TiO2. Nanoparticles of MoS2 3.0-4.5 nm in diameter were found to
catalyze the degradation of phenol, 4-chlorophenol, and pentachlorophenol under visible
4
light.[4, 7] The 8-10 nm size of MoS2 nanoparticle was not able to photodegrade phenol;
however, TiO2 sensitized with it was able to do so under visible light.[4] Both sizes of
MoS2 nanoparticle were noted to exhibit a blue-shift in absorbance due to quantum
confinement of charge carriers.[4,5] Wilcoxon et al. took great care to prepare sizeselected MoS2 nanoparticles to investigate size-dependent photochemistry. The laborious
synthesis was performed using inverse micelles with long-chain polyalcohol
surfactants.[5] No purification was done to remove the surfactant afterwards. It is not
clear if the presence of the surfactants affected the photocatalytic processes in solutions.
Degradation of organics by unsupported, nanosized WS2 was not studied by these authors.
MoS2 and WS2 were also used to prepare composite photocatalysts by depositing
them onto another photoactive semiconductor.[4,9] The presence of two different
semiconducting materials with mismatched bandgaps has a synergistic effect on the
efficiency of photocatalytic processes.[4,9] Electrons and holes photogenerated close to
the interface tend to migrate to the opposite sides driven by the difference in energies of
the valence and conduction bands (Figure 3). This decreases the chance of recombination
and increases the chance of photochemistry taking place on the outer surfaces.
5
hν
Conduction
ConductionBand
Band
e-
Reduction
Energy
h+
Oxidation
Valence
Valence Band
Band
Figure 3. Increase of the photocatalytic efficiency of composite semiconductors due to improved charge
separation across the interface.
MoS2 and WS2 grown on TiO2 were shown to sensitize it to visible light.[4,9] This
sensitivity extended the wavelength of light absorbed by TiO2 up to about 600 nm in the
case of WS2 or about 700 nm in the case of MoS2.[9] The resulting cocatalyst pairs were
able to catalyze the photodecomposition of methylene blue and 4-chlorophenol by visible
light relative to TiO2 alone.[9] The cocatalyst pair of MoS2/TiO2 was expected to
catalyze the formation of hydroxyl radical in water by two sets of reactions[9] :
1a. 2MoS2 (h+) + 2H2O → H2O 2 + 2H+
1b. H 2O2 → 2●OH
2a TiO2 (e-) + O2 → O2 2b. O 2- + H 2O → ●O2H + OH2c. ●O2H + H2O → H2O2 + ●OH
2d. H 2O2 → 2●OH
Note that the hole, h+, and the excited electron separated. The hole migrated to MoS2
from the junction of the two semiconductors while the electron migrated to TiO2. Each
6
reacts to produce hydroxyl radicals. In a sample of MoS2, WS2, or TiO2 alone, there
would be no junction between two semiconductors. So, the hole and the excited electron
would exist inside the same nanoparticle. Recombination of the charges is more likely
when h+ and e- are proximal, and interfacial redox processes are inhibited.
In a different study, Zong et al. achieved redox production of H2 from water by using
MoS2 particles deposited onto CdS .[10] Under irradiation with visible light
(wavelenghth > 420 nm), individual sulfides were almost inactive compared to the
composite material.
1.3. Carbonyl Based Synthesis of Nanosized MoS2 and WS2
Samples of nanosized MoS2 and WS2 can be prepared by precipitation in an organic
solvent.[11] This is a fairly simple one-pot synthesis that requires thermally
decomposing the metal carbonyl in the presence of dissolved sulfur. The precipitation is
achieved by first dissolving sulfur in the solvent by warming, followed by adding
Mo(CO)6 or W(CO)6 to the cooled solution and reheating it to 140°C.[11] p-Xylene has
been used previously as an organic solvent which refluxes at about 140°C. This
temperature is maintained for several hours to complete the reaction using molybdenum
or for several days for the use of tungsten.[11]
Decalin was substituted for p-xylene to improve the yield of WS2 because it had the
additional characteristic of being an alkane. This characteristic was desired to avoid
aromatic π-systems which might coordinate with tungsten. Coordination of the solvent
with tungsten was to be avoided because the proposed synthesis produces elemental
7
tungsten as an intermediate by decomposing the W(CO)6.[11] It was speculated that
trapping the tungsten in an intermediate stage by coordinating with the p-xylene solvent
may have been the cause of a lack of product for the first and second attempts at WS2
synthesis described in Chapter 2. Such a coordination compound between W(CO)3 and
p-xylene has been synthesized previously in the literature by heating W(CO)6 and
p-xylene in an organic solvent.[12-15]
Using decalin, the synthesis was then repeated in the presence of silica and titania
separately to deposit products on the inert or photoactive substrates. These modifications
are alterations to the process described in the literature.[11] Neither deposited nor
unsupported products of the precipitation in organic solvent have been previously studied
in the photodegradation of organics. Determining the photocatalytic properties of these
sulfides in the degradation of organics is the central purpose of this work.
8
1.4. Goals of the Reported Research
The goals of this research were:
1. Synthesis of MoS2 and WS2 by precipitation in an organic solvent without using
surfactants.
2. Synthesis of these sulfides in the presence of silica and titania by precipitation in an
organic solvent to form silica and titania deposited sulfides.
3. Characterization of the unsupported sulfides in terms of elemental composition and
size.
4. Verification of the presence of the quantum size effect for the unsupported sulfides.
5. Test the photocatalytic properties of each synthesized sulfide by use of the sulfide as
a catalyst in the degradation of methylene blue by light ≥ 400 nm.
6. Test unsupported sulfides for photocatalytic properties by use of the supported sulfide
as a catalyst in the degradation of acetone by light ≥ 400 nm.
9
Chapter 2. Synthesis of the Sulfide Photocatalysts
2.1. Synthesis of Unsupported MoS2 and WS2 Nanoparticles in p-Xylene
MoS2 and WS2 nanoparticles were synthesized similarly to the method described by
Duphil, Bastide and their colleagues.[11] All synthesis steps were carried out in a reflux
apparatus under argon purge to prevent reaction with oxygen. The reflux apparatus
consisted of a two-necked round bottom flask equipped with a reflux condenser. Valves
were inserted into the top of the reflux condenser and the second mouth through which
argon flowed. This way, the solvent could be degassed by allowing argon to flow from
the side mouth up and out through the top of the condenser with the condenser valve
removed during reflux. Positive argon pressure was maintained while transferring solids
to the flask through the side mouth by allowing argon to flow down the condenser from
its valve.
Before each synthesis, the glassware for reflux apparatus and the large stir bar was
washed each time, once already scrubbed visually clean, with wash grade acetone to
remove any grease. Then it was soaked in aqua regia for a minimum of two hours to
remove traces of metals. Finally it was rinsed with deionized water and dried with
acetone (Spectrum Chemical 99.5% pure). The exception was the synthesis of MoS2 in
p-xylene, which was rinsed with wash grade acetone. A sacrificial portion of reaction
solvent was refluxed inside the apparatus before each synthesis, cooled, and removed by
glass pipette as a final rinse.
Reactant amounts were often scaled up or down from the original paper, but the mole
to mole ratio of sulfur to metal carbonyl was kept at 2:1. For example, the reactant
quantities were often doubled resulting in a target of 4.6 x 10-4 moles of sulfur (Fisher
Scientific, sublimed) and 2.3 x 10-4 moles of metal carbonyl, Mo(CO)6 (Aldrich 98%) or
W(CO)6 (Strem Chemicals 99%), in 200 mL p-xylene (EM Science 98%) or decalin
(Aldrich 99%, anhydrous). Some synthesis runs were also scaled down. Additions of
solids to the reaction vessel were done under argon purge to prevent oxygenation of the
mixture and any grains of reactant stuck on the joint are rinsed down into the reaction
vessel by an aliquot of solvent from within. The actual measured amounts of reactants
and solvents for all sulfides synthesized alone can be found in Table 1.
Table 1. Reactant quantities for synthesis of unsupported metal sulfides in p-xylene.
Product Synthesis
MoS2
WS2
Metal Carbonyl Mass
(mg ± 0.1 mg)
60.0
81.0
40.0
Sulfur Mass
(mg ± 0.1 mg)
14.8
15.0
7.2
Solvent Volume (mL)
200
200
100
The sulfur was transferred to the p-xylene and warmed to reflux, 140-143°C, over
thirty minutes to both dissolve the sulfur and degas the solution. After being allowed to
cool to room temperature, the metal carbonyl is added, and the solution is heated again to
reflux. The reflux temperature is held for three hours for MoS2 or over four days for WS2,
and then the reaction vessel is allowed to cool to room temperature. The product mixture
is then distilled by heating to reflux under heavy argon purge to remove much of the
p-xylene. Finally the mixture is cooled under argon pressure and transferred into a brown
11
bottle.
Several color changers were observed for both of these reactions. The p-xylene itself
was clear, and remained clear after both the sulfur and metal carbonyl were dissolved.
Both carbonyls were white in solid form. When the temperature reached about 120°C,
the solution turned yellow. As the solution heated to 140°C, it became browner in color.
The solution then quickly turned very dark purple, but still looked reddish or yellowish
through thin portions of liquid. Solid is visible being stirred. After another half an hour,
the mixture looks almost black with a slight purple tinge. This color remains throughout
the rest of the synthesis, which appears thick with solid being stirred. When a sample of
this mixture is taken in a test tube, it is purple with a tinge of red.
When using the tungsten carbonyl, a yellow color appeared as the temperature
reached 140°C. The solution became darker yellow and slightly green in color over the
next five minutes and continued growing darker over the next hour becoming and slightly
brown. Overnight the solution developed a purple tinge. By the next day it was
brownish purple and growing darker, finally turning a brownish black. When thin layers
of the liquid are looked through, the solution looks red, yellow, or slightly orange
depending on the angle of the light. After being left to settle, the solution is still very
dark brown with red visible through slim portions of the liquid. There was no visible
solid, even after centrifuging. The solution never cleared or decolorized, but it was of a
dark orange color when viewed in a 1 cm cuvette. A slight film was noticed on the inside
of the round bottom flask which was a translucent orange. Some of this was scraped in a
p-xylene rinse with a glass stir rod and poured with the rest of the solution into a brown
bottle for storage. The synthesis was repeated with the same color changes observed and
12
produced no visible solid except for the slight film, which seemed to dissolve when
introduced to the rest of the solution.
2.2. Decomposition of W(CO)6 in the Absence of Sulfur
The synthesis of WS2 was repeated twice without the addition of sulfur for
comparison as the first and second control experiments listed in Table 2. The control
runs were also observed to turn from yellow, to greenish yellow soon after the carbonyl
was heated to a temperature of 140°C.
Table 2. Reactant quantities for control synthesis.
Metal Carbonyl Mass
Product Synthesis
(mg ± 0.1 mg)
44.9
Tungsten Control
20.3
Sulfur Mass
(mg ± 0.1 mg)
0
0
Solvent Volume (mL)
100
50
The solution stayed yellow for over an hour and was left overnight to reflux. The next
morning a reddish brown was observed, which slowly developed into a brown color and
finally a brownish black as before. At the meniscus the solution looked purple or orange
depending on the angle of the light. In the 1 cm cuvette the solution looked orange like
the synthesis of WS2 solution. There was no solid visible, even after centrifuging. The
control was repeated a second time with the same results, but there was also an orange
film noticed on the inside of the flask. The fact that the bulk of the solution, which did
not have sulfur added, developed the same color as the solution containing sulfur as a
reagent, means that the color of the solution is not due to WS2 formation. One
explanation for the lack of formation of solid during the WS2 synthesis in p-xylene would
be that the tungsten was trapped by a side reaction. A likely candidate for this side
13
reaction is the formation of a π-complex such as a “piano stool” (η6-C8H10)W(CO)3,
which has been synthesized by heating W(CO)6 to 135°C in the presence of p-xylene and
by other methods.[12-15]
2.3. Synthesis of Unsupported MoS2 and WS2 Nanoparticles in Decalin
Decalin was chosen as a replacement for the solvent p-xylene because of its lack of
the aromatic double bounds potentially used to form coordination bonds with tungsten or
W(CO)3. Synthesis was then repeated for both WS2 and MoS2 using decalin. It should
be noted that temperature fluctuations of 140-155°C using decalin as a solvent were
higher than those observed with p-xylene, but no problems resulted. The exact reactant
quantities are listed in Table 3.
Table 3. Reactant quantities for synthesis of unsupported metal sulfides in decalin.
Metal Carbonyl Mass
Sulfur Mass
Product Synthesis
Solvent Volume (mL)
(mg ± 0.1 mg)
(mg ± 0.1 mg)
MoS2
15.5
3.6
50
WS2
40.5
7.4
100
During the synthesis of WS2 in decalin, it was noticed that the decalin solution
remained clear, even after the addition and dissolution of sulfur. Once the carbonyl was
added and dissolved with the temperature reaching 150°C, the solution was a slight gold
color. The metal carbonyl itself had been white. After about twenty minutes at this
target temperature, the color has become more beige, and then it acquired a slightly red
tinge. This color remained for several hours, and the reaction was allowed to continue
over the weekend. After two days, solid had begun to settle to the bottom in spite of
stirring and adhered to the inside of the flask. The mixture and the solid were a rusty
14
brown color.
Observations recorded during the synthesis of MoS2 in decalin were that the decalin
solution remained clear throughout the dissolution of sulfur until the carbonyl had been
dissolved and the temperature reached 110°C. At that point the solution turned from
colorless to a brownish color. Over ten minutes the temperature was raised to 139°C and
the solution became redder and rust colored. The solution then quickly turned nearly
black, but still looked brown through thin portions of the liquid. There are some dark
flecks of solid being stirred. After half an hour since the solution reached 139°C, the
solvent looks almost green, and there is enough floating solid to see that it is a reddish
brown color. After another ten minutes, it is more uniformly black, but fifteen minutes
later the edges of the flask look more of a green or brown color. After yet another fifteen
minutes the solution still looks black but with a purple tinge and stays that way
throughout the last hour of heating. When a sample of this mixture is taken in a test tube,
it looks purple with a red tinge. This is a similar appearance to the MoS2 mixture
synthesized in p-xylene.
After concentrating under argon purge, then cooling, synthesized sulfides were then
isolated from their solvents. This required several steps. Each mixture was agitated and
transferred to a polysulfone bottle. These were ultacentrifuged at 5000 rpm for fifteen
minutes, and the clear, colorless supernatant was carefully removed by pipette. About
100 mL of cyclohexane was added. Nanoparticles were re-dispersed by shaking, then
sonication for several seconds each. This cycle of ultracentrifuging and re-dispersion
was repeated four more times. The samples were then dried in a rotary evaporator. The
dry nanoparticles of MoS2 and WS2 which were synthesized in decalin are displayed in
15
MoS2
WS2
Figure 4. Dried products from the synthesis of unsupported MoS2 and WS2.
Figure 4. Once dry, the nanoparticles were re-dispersed readily in chloroform and stored
in brown bottles for future use.
2.4. Synthesis of Silica Deposited MoS2 and WS2 Nanoparticles
Synthesis of both metal sulfides was also carried out in the presence of silica (Aldrich,
99.5% 10-20 nm nanoparticles) to form deposited catalysts. For this purpose, several
modifications were made to the procedure. First, a mixture of silica in decalin was
heated to reflux, 170-180°C, over an hour’s time with argon purge to degas the decalin
and drive off any moisture from the silica. Afterwards, it was cooled back to about 40°C
before adding the sulfur and continuing the procedure as before. Table 4 lists the
amounts of reactants used along with the expected mass ratio of silica to sulfide with the
16
metal carbonyl as the limiting reagent.
Table 4. Reactant quantities for synthesis of silica supported metal sulfides.
Metal Carbonyl
Sulfur Mass
Solvent
Silica
Silica to Sulfide
Product
Mass
(mg ± 0.1 mg) Volume (mL) Mass (g)
Mass Ratio
Synthesis
(mg ± 0.1 mg)
MoS2 / SiO2
30.5
7.4
100
1.80
97.3 : 1
WS2 / SiO2
40.3
7.5
100
2.85
100. : 1
These reactions go through several color changes. The silica in hot decalin looks
more like a sludge or gel than a powder, and this does not change after sulfur is dissolved.
For MoS2 synthesis, the mixture turns brown once the carbonyl is dissolved and the
temperature reaches about 120°C. Then it quickly takes on a greenish hue. After quickly
increasing the temperature to 140°C over two minutes, the color has changed to a very
dark brown and stays there. Once the mixture has cooled and the silica has settled, it
appears that the silica itself is brown throughout while the solvent is clear. Even under
light stirring, the brown color does not seem to spread into solution but sticks with silica.
During WS2 synthesis with silica, the mixture turns an amber or golden color after the
carbonyl is dissolved and the temperature reaches about 125° C. Five minutes later the
temperature had reached 150°C and the color was the same. After another thirty minutes,
the color was slightly more brownish. Two hours later, the color is best described as
butterscotch, which became slightly darker over the next hour. After heating overnight,
the silica mixture looks uniformly rusty brown. This color remained throughout the day.
By the third day, the color had changed to brown without any red tinge. Letting the silica
settle to the bottom leaves the decalin clear and the silica uniformly tan colored.
Each silica deposited sulfide was then removed from decalin as described for the
removal of free nanoparticles from their solvents. Then each was dried by rotary
17
evaporation. Instead of being re-dispersed in chloroform, the silica deposited sulfides
were dried, kept in a bottle, and used as needed.
2.5. Synthesis of Titania Deposited MoS2 and WS2 Nanoparticles
Titania deposited metal sulfides were synthesized by replacing the silica in this
procedure with titania (Aldrich, 99.7% pure anatase nanoparticles less than 25 nm
diameter). The anatase form of TiO2 which was used is one of the crystalline
modifications along with rutile and brookite. For MoS2, the synthesis was let to heat over
the weekend as an additional modification to ensure the reaction completed. The exact
amounts of reactants used are listed below in Table 5.
Table 5. Reactant quantities for synthesis of titania supported metal sulfides.
Metal Carbonyl
Sulfur Mass
Solvent
Titania
Titania to Sulfide
Product
Mass
(mg ± 0.1 mg)
Volume (mL) Mass (g)
Mass Ratio
Synthesis
(mg ± 0.1 mg)
MoS2/TiO2
30.2
7.4
100
1.84
100. : 1
WS2/TiO2
40.2
7.5
100
2.85
100. : 1
During heating, titania suspension in decalin had the appearance of milk, however
dissolving the sulfur gave it a tan or peach color that appeared homogeneous. This stands
out because dissolving sulfur in solvent alone or solvent plus silica did not result in a
color change. When using MoS2, this color remained until the mixture reached 140°C.
At that time, it was noticed that the mixture had become darker in color. The tan color
shifted to a light brown through to the end of the synthesis. For WS2 synthesized with
titania, there was no noticeable change from the peach color over a period of several
hours. Left overnight, the mixture became perceptively darker to a light brown. For both
sulfides, the brown color appears to be from the solid and to be uniform, while the
18
solvent left behind is clear.
Each titania deposited sulfide was then removed from decalin and dried. This process
consisted agitating each mixture and transferring about 20 mL of it into a polypropylene
bottle. Using the smaller, polypropylene bottle was a change in procedure for
convenience in isolating nanoparticles from small volumes of decalin. The mixture was
agitated and then centrifuged for 10 minutes at 12,000 rpm. The clear supernatant was
carefully removed by pipette. Then, about 20 mL of acetone (Spectrum Chemical 99.5%)
were added and the bottle was shaken to re-disperse the pellet. Acetone could safely be
used with these bottles, so it was used instead of hexane. This process of centrifuging
and re-dispersing was repeated three more times. The nanoparticles were then dried in
the rotary evaporator as before. Redispersion of silica and titania supported catalysts was
accomplished by shaking rather than sonication to avoid the detachment of supported
MoS2 and WS2.
19
Chapter 3. Characterization of Synthesized
Nanoparticles
3.1. Imaging of Nanoparticles with Transmission Electron Microscopy
In order to observe the nanoparticles and verify the size, images were taken of the
product mixture from each synthesis with a transmission electron microscope (Electron
Microscopy Sciences, part No. I400-Cu). The sample grid for the transmission electron
microscope, TEM, was a carbon film supported by a copper grid. Sample grid
preparation consisted of agitating the product mixture, removing less than a milliliter
with a disposable glass pipette, and drying drops of product mixture on the sample grid.
TEM images of the MoS2 synthesized in p-xylene reveal nanoparticles ranging from
about 5 nm to 30 nm in diameter (Figure 5).
The very small specs in the image alongside the larger, dark bodies illustrate the wide
range of nanoparticle sizes formed. This collection of relatively small nanoparticles was
located near a large agglomeration visible on the right and top side of the photo.
Figure 5. TEM Image of MoS2 in p-xylene at 60,000x magnification.
Collections of individual nanoparticles were often observed near larger
agglomerations. It is not clear from the images whether these agglomerations consist of
separate particles that are physically in contact with each other, or bound together by a
lattice into a large contiguous structure.
The sizes of particles formed in either p-xylene or decalin were determined by
comparing the TEM images of products with a scale derived from a TEM image of a
catalase standard (Electron Microscopy Sciences, part No. 80014) (see Figure 6).
21
Figure 6. TEM image of the Catalase standard at 200,000x magnification.
Two-dimensional packing of the catalase molecules results in a grid of lines 6.85 nm
x 8.75 nm. The measured distance between the lines provides a scale. Figure 6 shows
the negative image of the catalase standard with some of the lines accented in Powerpoint
to make them clearer to see. The scale drawn from these regularly spaced lines can be
adjusted depending on the magnification. The bright white specs are adventitious
particles like lint, which the scanner commonly detects in TEM negatives. Any slight
creases or bends in the film can also be seen, like the slight curve running from top to
bottom near the center of this image.
Figure 7, another TEM image of the MoS2 synthesized in p-xylene with a larger
22
magnification verifies the presence of nanoparticles with diameters of 5-25 nm.
Figure 7. TEM image of MoS2 in p-xylene at 150,000x magnification.
There appear to be many nanoparticles in this photo with diameters of about 10 nm or
less using the scale on the left side and several particle sizes have been labeled for
comparison.
As noted earlier in the synthesis section, the WS2 synthesis in p-xylene did not yield
any visible product, even after use of the centrifuge. This was likely due to the formation
of a side product with p-xylene.[12-15] TEM revealed that the solution did contain very
few particles. These were shaped like tissues like the one in Figure 8, and did not
resemble those documented by previous researchers.[11] The close-up image of one
23
tissue shows the folds in the structure.
Figure 8. TEM image of WS2 in p-xylene at 100,000x magnification.
This may be a thin layer of nanosized WS2 which has begun to fold in on itself, or an
unintended side product. The approximate size of the tissue in the folded form appears to
be 300 nm × 750 nm, but the folds prevent the exact size of the sheet from being
determined by TEM.
Decalin was chosen as a replacement for the solvent p-xylene because of its lack of
double bonds or aromaticity. These properties were desired to avoid the potential
formation of coordination compounds with tungsten or tungsten carbonyl. MoS2 was
synthesized in decalin to check if sulfide nanoparticles would form in this new solvent.
24
The new particles were imaged in Figure 9. Drying the sample on the TEM plate resulted
in the particles grouping together, yet retaining their individual boundaries. The
boundaries can be seen in Figure 9 where the particles form peninsulas off the main
group because fewer particles overlap. Reoccurring particle diameters represented in this
photo were about 15 nm, 8 nm, and 5 nm. It is unclear from the relatively low resolution
of these TEM images whether the MoS2 nanoparticles are amorphous or crystalline;
however, Duphil et al. [11] found that the nanoparticles tend to be amorphous under
similar reaction conditions.
Figure 9. TEM image of MoS2 in decalin at 200,000x magnification.
TEM images of WS2 in decalin reveal particles as small as about 5nm in diameter
25
(Figure 10). Although a few particles as small as 5 nm in diameter were found in this
image, most were about 10nm in size. Individual particle sizes can best be seen near the
edges of the group of particles, much like in the MoS2 samples. Figure 11 displays an
image of one such area from a WS2 synthesis in decalin with greatly enhanced contrast.
Figure 10. TEM image of WS2 in decalin at 200,000x magnification.
26
200,000x
100nm
5nm
Figure 11. TEM image of WS2 in decalin at 200,000x magnification with high contrast added.
In this image, nanoparticles about 5 nm in size were found surrounding a large peninsula
of an agglomeration. This image confirms the formation of WS2 nanoparticles under 10
nm in diameter.
The WS2 nanoparticles synthesized in decalin are comparable in size and appearance
to those found in the samples of MoS2. In scanning across the sample, a majority of the
nanoparticles with visible borders appear to be about 15 nm in diameter (Figure 12).
There are also particles in the 20-30 nm diameter size similar to the particles observed for
MoS2. These results are too qualitative to determine if the WS2 synthesis produced
routinely smaller or larger particles than the MoS2 syntheses.
27
Figure 12. TEM image of WS2 in decalin at 100,000x magnification.
3.2. UV-visible Spectra of the Nanoparticles
The UV-visible absorption threshold can allow estimation of the electronic bandgap
in the sample. In addition, UV-visible spectra of particles of different sizes can verify if
the quantum confinement of charge carriers takes place. Within the quantum
confinement regime, as the particle size decreases, the bandgap value increases as fewer
molecules contribute their d-orbitals to the valence and conductance bands.[6] This leads
to a blue-shift in the absorption threshold of the nanoparticles.
Samples of MoS2 synthesized in p-xylene and decalin as well as WS2 synthesized in
28
decalin were tested for nanoparticles exhibiting a blue-shift in absorbance. The Agilent
diode array UV-visible spectrometer model 8453 was used to record the UV-visible
spectra of the synthesized samples. Each of the samples was agitated and then let to
settle in a 1 cm quartz cuvette as spectra were taken over time. Large particles are
expected to settle first, while smaller particles stay suspended for longer periods of time.
A shift in the size distribution of suspended particles towards small sizes exhibiting
quantum confinement would be expected to cause a shift in the resulting absorption
spectrum. By subtracting a spectrum from another taken earlier, the spectrum of the
particles which settled to the bottom in the meantime can be obtained as a difference
spectrum.
Figure 13 displays difference spectra obtained during sedimentation of MoS2 from the
batch synthesized in p-xylene. Sedimentation is carried out in p-xylene and monitored by
taking UV-visible spectra over time. The reference spectrum of p-xylene was taken
before the start of sedimentation. For comparison, these difference spectra have been
normalized to a peak absorbance of 1.
29
Difference Spectra of Molybdenum Sulfide Settling in p-Xylene
Normalized Absorbance (arb. units)
1.1
particle size decreases
1
0.9
0.8
0.7
0.6
0.5
Particles sedimented
within 300-900 min
within 90-300 min
within 30-90 min
0.4
0.3
0.2
300
400
500
600
700
800
900
1000
1100
Wavelength (nm)
Figure 13. UV-visible difference spectra determined from spectra obtained during the settling of MoS2 in
p-xylene.
It is easy to see that particles settling out within 30-90 min. have an absorption peak
around 600 to 700 nm. Compared to them, smaller particles which settled out within 90300 min. exhibit a blue-shift in absorbance peaking at about 500 nm. Even smaller
particles settling within 300-900 minutes exhibit an even stronger blue-shift in
absorbance peaking at about 430 nm.
Their spectrum also begins to exhibit a sharper absorption threshold similar to the 4.5
nm particles studied by Wilcoxon.[4] This further indicates the presence of particles
exhibiting quantum confinement. The long tail to this threshold could be due to the
remainder of larger particles which had not settled until this time.
The blue-shift in absorbance is also observed for WS2 synthesized in decalin (Figure
14). Particles settling out within the first twenty minutes are mostly large and absorb
30
throughout the visible spectrum. In the next forty minutes, the settling particles
contributed more to the absorbance of light between 700-800 nm. The next set of
particles favored even shorter wavelengths of light. During the following night, the
distribution of particles settling had a peak absorbance at about 550nm. The centrifuge
managed to remove nanoparticles with a peak absorbance around 500nm.
Difference Spectra of Tungsten Sulfide Settling in Decalin
1.1
Normalized Absorbance (arb. units)
particle size decreases
1
0.9
0.8
0.7
0.6
0.5
0.4
300
Particles sedimented
within 138-1067min
within 69-138 min
within 18-69 min
by centrifuge
400
500
600
700
800
900
1000
1100
Wavelength (nm)
Figure 14. UV-visible difference spectra determined from spectra obtained during the settling of WS2 in
decalin.
MoS2 synthesized in decalin was allowed to settle in decalin and monitored to check
if the size-dependent blue-shift was present (Figure 15).
31
Difference Spectra of Molybdenum Sulfide Settling in Decalin
Normalized Absorbance (arb. Units)
0-20 min
20-63 min
1
63-120 min
0.8
15 per. Mov. Avg. (120190 min)
0.6
0.4
0.2
0
300
400
500
600
700
800
900
1000
1100
Wavelength (nm)
Figure 15. UV-visible difference spectra determined from spectra obtained during the settling of MoS2 in
decalin.
The blue-shift is certainly evident when comparing the spectra of the particles settling out
during these three time periods, the first twenty minutes, the next forty-three minutes, and
then the final fifty-seven minutes. An absorption threshold seems to be forming. The
spectra for the next sixty minutes became complex however, and even a fifteen point
moving average still has many peaks instead of a threshold.
3.3. Energy Dispersive X-Ray Analysis
Energy dispersive X-ray (EDX) analysis is an elemental analysis method. This
method is based on the emission of characteristic X-ray photons by an atom after it has
been excited by an electron beam impact. The analysis is performed with a scanning
32
electron microscope (SEM JEOL 6300 with Noran EDX adapter). The presence of both
tungsten and sulfur was determined in a sample of WS2 which had been synthesized in
decalin (Figure 16). This determination was done using an area of 1 µm × 1 µm, and so
the results reflect the average composition of a group of nanoparticles.
Figure 16. Energy dispersive X-ray analysis of WS2 synthesized in decalin.
The peaks at 0.3 keV, 1.4 keV, 1.9 keV, and 2.1 keV indicate the presence of tungsten
and are labled with a “W” in the figure. Peaks at 0.6 keV and 2.3 keV confirm the
presence of sulfur and are appropriately labeled with an “S.” Copper peaks at 0.9 keV,
33
8.0 keV, and 8.9 keV marked with a “Cu” are from the sample grid.
The first labeled peak corresponds to both carbon from the sample grid and tungsten, so it
is also labeled with a “C.” Peaks attributable to tungsten and sulfur from the EDX
analysis indicate that WS2 was formed.
Performing EDX analysis on a sample from the MoS2 synthesized in decalin detected
a carbon peak from the grid at 0.3 keV, copper peaks at 0.9 keV and 8 keV, and a peak at
2.4 keV where both molybdenum and sulfur respond (Figure 17).
Figure 17. Energy dispersive X-ray analysis of MoS2 synthesized in decalin.
34
Separate molybdenum and sulfur peaks at 2 keV and 0.6 keV respectively confirm the
presence of both molybdenum and sulfur. This would indicate that MoS2 was formed.
35
Chapter 4. Photocatalytic Degradation Tests: Setup and
Protocols
4.1. Photocatalysis Setup
The light source is a 300 W xenon arc bulb in a fan-cooled, universal housing from
Newport powered by the Newport power supply (model 69911) that is set to the
maximum output of 300 W in constant power mode. Light from the source first passes
through a focusing lens and a water-cooled water filter that removes infrared radiation.
Then, it passes through a selected long-pass filter (Edmund Industrial Optics) that
removes high energy UV light and only transmits UV-visible radiation with the
wavelength above a certain threshold. Then a mirror reflects the beam straight down into
a beaker containing the solution of a compound to be degraded and the catalyst (Figure
18). The path of the light can be traced through each attachment in Figure 19.
Figure 18. Close-up view of the UV-visible light beam irradiating the beaker with a reaction mixture.
Long-pass cut-off filter is visible in the upper left corner.
La m p H o usin g
W ate r F ilte r
Le n s
W ate r
C ircu la tor
S h u tte r
M irror
L igh t Pa th
S hu tter
Co n trol
E vap o ratin g
d ish
M a g ne tic
S tirrer
P ow e r S u pp ly
L o ng P ass Filte r
Figure 19. The path of the light through the experiment setup.
37
The mixture is stirred rapidly throughout the experiment by a magnetic stirrer, and is
covered by a watch glass to slow evaporation of the solution. Both watch glass and
beaker are pyrex, which has an absorption cutoff around 310 nm. This cutoff is far lower
in wavelength than any of the long-pass filters used and therefore does not affect the
incident light spectrum. The lens, water filter, long-pass filter holder, and reflecting
mirror are all attachments linked together to the lamp housing by screws, so the light path
up to this point is constant in length. The beaker and watch glass are always re-centered
under the beam after a sample is taken. The condenser is adjusted to narrow the beam as
much as possible. Under these conditions, the light beam is slightly diverging and has a
radius slightly larger than the 250 mL beaker. The beaker itself is placed inside of an
evaporating dish covered in reflective foil to maximize the exposure of the beaker to the
light. About 20 mL of Millipore water are added to the evaporating dish before each run
to dampen any temperature fluctuations. This water did not warm above room
temperature, indicating that the water filter was successful in removing the IR component
of the lamp spectrum.
4.2. Photocatalytic Tests Protocols
Photodegradation of methylene blue was used to test the photocatalytic activity of six
synthesized catalysts, namely, unsupported MoS2 and WS2 nanoparticles synthesized in
decalin, MoS2 and WS2 supported on SiO2, and MoS2 and WS2 supported on TiO2
(anatase). Solutions of methylene blue were irradiated in the presence of each of the six
synthesized catalysts with light of wavelength ≥ 400 nm (400 nm long-pass filter). The
38
absorbance peak at 665 nm was used to monitor the degradation of methylene blue with
the Agilent 8453 UV-visible diode array spectrometer. Unsupported MoS2 and WS2
nanoparticles were also tested in degradation of acetone. Acetone solutions were
irradiated using the 400 nm long-pass filter. The absorbance peak at 265 nm was used to
monitor the degradation of acetone with the HP 8452A UV-visible diode array
spectrometer. To account for autodegradation of methylene blue, acetone evaporation,
and any other unforeseen effects, control experiments were performed, in which metal
sulfide catalysts were absent.
4.2.1. Methylene Blue Degradation by Unsupported MoS2 and WS2
In preparation for testing un-supported catalysts, the MoS2 or WS2 is first dispersed in
chloroform and then dried in a clean, dry 250 mL beaker, coating the bottom with a
visible layer. The beaker for the control trial is not coated. A 200 mL solution of
methylene blue with is added to each beaker by graduated cylinder. The methylene blue
solution is then stirred vigorously with a magnetic stir rod for at least half an hour prior to
taking any samples for absorption measurements. The mixture is not irradiated during
this period. Stirring continues throughout the trial. Spectra of three separate samples of
the 200 mL reaction mixture are taken to determine an average original absorbance. The
three absorptions are checked to ensure they are within a range of 0.02 abs units.
Unsupported MoS2 and WS2 nanoparticles exhibit hydrophobic properties and do not
disperse in aqueous solution. During the experiment, they remain attached to the bottom
of the beaker or deposit onto a Teflon stirrer.
The standard operating procedure for taking a sample has several steps. A
39
background spectrum is obtained with Millipore water. Then the quartz cuvette is rinsed
with less than half a milliliter of reaction mixture using a disposable glass pipette, and
finally that pipette is used to deliver about 1 milliliter of reaction mixture to the cuvette.
The spectrum is taken as soon as possible with an integration time of 5 seconds. When
the cuvette is not in use, it is rinsed with and let to soak in Millipore water. The lamp is
warmed up for at least 15 minutes with the shutter closed before it is used to irradiate the
reaction mixture. Finally the shutter is opened. To monitor the absorbance of methylene
blue, samples are taken throughout the next 6 to 9 hours and a UV-visible spectrum is
taken of each sample. Absorption measurements taken from the beakers containing
MoS2, the control, and WS2 are used as data for Trials 1, 2, and 3 respectively.
4.2.2. Methylene Blue Degradation by MoS2 and WS2 Supported on SiO2 or TiO2-Anatase
Methylene blue solutions for photocatalytic testing of deposited sulfides were
prepared through a series of steps. First, a stock solution of 6.25 x 10-5 M methylene blue
in Millipore water was prepared using 95% pure methylene blue hydrate (MW = 319.85),
10. mg in a 500.0 mL volumetric flask. Then a clean, dry 100 mL beaker was used to
mass the amount of solid listed in Table 6for the sulfide or control being tested.
Table 6. Quantities of solid used as catalysts to test deposited sulfides.
Solid Tested
Mass (mg ± 2 mg)
WS2/SiO2
16
MoS2/SiO2
14
SiO2
14
WS2/TiO2
15
MoS2/TiO2
15
TiO2
15
Of the stock solution, 25.0 mL was added along with 60.0 mL of Millipore water by
graduated cylinder to the 100 mL beaker followed by a magnetic stir rod. The mixture
40
was then allowed to stir vigorously for two hours to disperse the solids. Three samples
were taken shortly before the lamp shutter was opened. Samples were taken by standard
operating procedure, with the exception that the cuvette was also sonicated while not in
use. The lamp was again warmed up for at least fifteen minutes with the shutter closed
before use. The lamp setting of 300 watts constant power remained but the filter was
changed to a 400 nm long pass filter. Reaction mixtures were irradiated for about 8 hours
for the silica-deposited sulfides and about 4 hours for the titania-deposited sulfides.
Throughout this time, samples were taken for UV-spectroscopy as before.
4.2.3. Accounting For Light Scattering by SiO2 and TiO2
SiO2 and TiO2 have hydroxylated wettable surfaces. These powders easily disperse
in aqueous solutions even if MoS2 or WS2 are deposited on them. Therefore, the
catalysts in this case are not attached to the bottom of the beaker, as was in the case of
unsupported MoS2 or WS2. The photocatalyst is expected to be more efficient if
dispersed throughout the mixture.
Highly dispersed SiO2 and TiO2 cause sufficient light scattering that manifests as
a sloping baseline in the UV-visible spectra. This sloping baseline is superimposed onto
the absorption peak of methylene blue at 665 nm. This effect can be seen in the case of
silica deposited WS2, for example (Figure 20). A UV-visible spectrum of a 200 mL
solution of methylene blue was taken before stirring in 29 mg of silica deposited WS2 for
1.5 hours and taking another spectrum. The mixture has an absorption measurement
between 360 nm and 440 nm due to scattering, and the absorption at 665 nm decreased
substantially due to adsorption of methylene blue onto the silica.
41
0.9
Absorbance (arb. units)
0.8
A
0.7
Methylene Blue
0.6
0.5
0.4
0.3
0.2
0.1
0
200
300
400
500
600
700
800
Wavelength (nm)
0.9
Absorbance (arb. units)
0.8
0.7
0.6
0.5
Methylene Blue
and WS2/SiO2
0.4
Methylene Blue
Absorption
0.3
0.2
0.1
0
200
Light
Scattering
B
Linear Correlation
300
400
500
600
Light
Scattering
700
800
Wavelength (nm)
Figure 20. (A) UV-visible spectrum of aqueous methylene blue. (B) UV-visible spectrum of aqueous
methylene blue in the presence of WS2/SiO2. Light scattering by silica offsets the absorption peak of
methylene blue. The offset can be estimated using the linear correlation with the light scattering
contribution in a remote part of the spectrum far from the absorption peaks.
42
Taking further spectra over time, it was also observed that increases in the
scattering due to more fully dispersing the silica are reflected in an increase in the 665 nm
absorption measurement (Figure 21).
The Changing Spectrum of Methylene Blue and Silica Deposited
Tungsten Sulfide after Minutes of Stirring
Absorbance (arb. units)
1.2
1
310 minutes
0.8
147 minutes
120 minutes
0.6
114 minutes
97 minutes
0.4
90 minutes
0.2
0
200
300
400
500
600
700
800
Wavelength (nm)
Figure 21. An increase in scattering by the dispersing catalyst contributes an observable increase in the
665 nm absorbance measurement as the mixture stirs over time.
As the time spent stirring increases, the catalyst becomes better dispersed, which results
in an increase in scattering. This increase can be seen clearly in the range between 360
nm and 420 nm and coincides with an increase in the absorbance of the methylene blue
spectrum.
In order to determine the absorbance due to methylene blue, the scatter contribution
under the peak at 665 nm can be estimated and subtracted. For this, one can start with a
solution that contains no methylene blue and only the scattering solid (SiO2 or TiO2).
Then one can determine the scatter level at 665 nm and in the region far from the
absorption signatures. As shown in the next chapter, for each solid photocatalyst, there is
43
an empirical proportionality between the two values and it can be conveniently exploited.
By measuring the scatter level far from absorption peaks, one would use the
proportionality to reconstruct the scatter contribution at 665 nm under the methylene blue
absorption peak. This contribution will have to be subtracted from the peak absorbance
to yield the correct methylene blue signal. Note that calculating and subtracting the
scattering must be done individually for each spectrum.
The remote spectral region far from absorption peaks was chosen at 375 nm (for
TiO2-containing catalysts) or 420 nm (for SiO2-containing catalysts). 15 mg of each
deposited sulfide as well as titania and silica were each added to 85.0 mL of Millipore
water in a 100 mL beaker. The measured masses are listed in Table 7 below.
Table 7. Quantities of solid used in tests to determine proportionate scattering.
Solid Tested
Mass (mg ± 2 mg)
WS2/SiO2
15
MoS2/SiO2
15
SiO2
14
WS2/TiO2
15
MoS2/TiO2
15
TiO2
15
The silica, titania, and deposited sulfide mixtures were stirred vigorously for over two
hours before samples were taken for UV-spectroscopy. Samples were taken from each
mixture over a period of several hours. The measured absorption at 665 nm was plotted
versus absorption at 420 nm (for SiO2 containing mixtures) or versus absorption at 375
nm (for TiO2 containing mixtures). A satisfactory linear relationship was found in each
case as shown in the next chapter.
44
4.2.4. Acetone Methylene Blue Degradation By Unsupported MoS2 and WS2
Rates of degradation of acetone with unsupported WS2 or MoS2 were determined. As
a control, the degradation rate of acetone in absence of photocatalysts was determined as
well. A 0.02 M solution of acetone in Millipore water was first prepared. Then, 50.0 mL
of .02 M acetone was added to three 250 mL beakers. One was clean and dry for the
control trial, the other two had WS2 or MoS2 dispersed in chloroform dried in them. The
masses were 18 mg of WS2 or 11 mg of MoS2. Drying was done immediately before
weighing and then adding the acetone solution followed by 150.0 mL of Millipore water.
The mixture was then stirred for at least fifteen minutes. The same lamp setup was used
as for the silica-deposited sulfide trials as well as the titania-deposited sulfide trials, and
the sulfide appeared to adhere to the bottom of the beaker and stirrer as with the free
nanoparticles. The magnetic stirrer in this case was set to two out of 10 to slowly stir the
mixture. One sample was taken directly before opening the lamp shutter and the
absorption of this at 265 nm was used as A0. During 4 to 7 hours of irradiation, samples
were taken as before for UV-visible spectroscopy to monitor the degradation of acetone
by relative absorption.
45
Chapter 5. Photocatalytic Degradation Tests: Results
5.1. Methylene blue degradation by unsupported MoS2 and WS2
During the degradation of methylene blue by a photocatalyst, the reaction can be
monitored by observing a prominent absorption peak of methylene blue at 665 nm. If the
original absorbance at 665 nm is A0, then it decreases to A as methylene blue is degraded.
Figure 22 shows the methylene blue photodegradation kinetics under visible light.
Degradation of Methyene Blue with a Long-Pass 435 nm
Filter by Metal Sulfide Catalysts
1.2
1.1
A/A0
1
0.9
0.8
Molybdenum Sulfide
0.7
Tungsten Sulfide
Control
0.6
0.5
0
2
4
6
8
10
Time (Hrs)
Figure 22. Relative UV-visible absorption at 665 nm plotted over time using unsupported MoS2 and WS2
as the catalysts.
The relative absorption, A/A0, provides a convenient measure of the extent of reaction. It
has to be noted that the original absorbance, A0, is usually an average of measured
several measurements before the photoreaction is started.
Three separate trials were conducted: on unsupported nanoparticulate MoS2, WS2,
and in the control experiment with no photocatalyst added. Methylene blue concentration
decreases in all three cases. This means that in addition to the sulfide-catalyzed
heterogeneous process, there is also a possible self-degradation of methylene blue due to
the interaction of the optically excited dye with dissolved oxygen in the solution. The
exact mathematical expression for the reaction kinetics is not known.
If the reaction rate is proportional to the concentration of methylene blue, the firstorder kinetics would lead to an exponential decay curve. Zero-order kinetics could lead
to linear decay plots if the reaction rate is independent of the concentration of methylene
blue in the solution. This can be the case if the rate-limiting step is the diffusion of
photoproduced electrons and holes to the surface or in the case of some adsorptiondesorption equilibria.[16] Over the period of 6-8 hours, the curvature in the kinetic data
is not apparent. On this time scale, each data set can be approximated with a line. Firstorder exponential curves did not produce significantly better fits as judged by r-squared
values.
Regression analysis was performed in Microsoft Excel to find the best-fit line for
each trial, and the obtained slopes and fitting errors are listed in Table 8. Comparisons of
the reaction rates are made by comparing the slopes of the fits. Statistical analysis was
carried out focusing on the slope for each trial to determine if the different slopes were a
result of random error. The standard deviation of each slope was determined using the
47
LINEST function in Excel. The margin of error for each slope was calculated at the 90%
confidence level (±1.645 δ, where δ is the standard deviation).
Table 8. Analysis of the rates of the methylene blue degradation by unsupported MoS2 and WS2. Slopes
of the linear fits from Figure 22 are listed along with the error limits at the 90% confidence level.
WS2
Control (No Catalyst)
MoS2
R2 Value
0.9979
0.9926
0.9976
-2
-1
Slope (x10 hr )
-3.7 ± 0.1
-4.6 ± 0.296
-4.13 ± 0.15
Upper Limit (x10-2 hr-1)
-3.6
-4.4
-3.98
Lower Limit (x10-2 hr-1)
-3.8
-4.9
-4.28
In the presence of WS2, the relative absorption at 665 nm decreases faster over time
compared to the control. The slope of the plot is more negative, indicating that the
degradation of methylene blue was catalyzed by WS2 nanoparticles. At the same time,
MoS2 failed to catalyze the degradation of methylene blue under the reaction conditions.
Moreover, nanosized MoS2 apparently exhibits a less negative slope than the control,
which lacked any catalyst. This indicates the use of MoS2 led to less photodegradation of
methylene blue than the photobleaching observed without catalyst. This underscores the
challenges of photocatalytic measurements under conditions when the photocatalyst is
not uniformly dispersed in the solution but remains attached to the bottom of the reactor.
Some of these issues relate to the experiment design. Deposition of the
photocatalysts by drying the chloroform suspensions is inherently not very reproducible.
It can introduce variations of the solid layer thickness from experiment to experiment. In
addition, when a Teflon stirring bar spins on the deposited nanoparticulate layer, the
nanoparticles are capable of transferring from the glass onto the Teflon stirring bar over
the course of an experiment. This introduces another variable into the kinetics. These
issues can be overcome by “solubilizing” the photocatalyst in the reaction mixture by
means of depositing it onto a highly dispersed hydrophilic substrate. This strategy is
48
explored by using highly dispersed SiO2 and TiO2 powders as described in the following
sections.
Other considerations involve the choice of an organic compound for
photodegradation. Since the MoS2 and WS2 nanoparticles are very hydrophobic, they
likely undergo clustering in aqueous media, which decreases the sulfide-water interface
area. Interfacial charge transfer of electrons and holes would be hindered and, logically,
this would greatly diminish the photocatalytic efficiency. The overall kinetics might
become dominated by side reactions, such as self-degradation of a dye due to electronic
excitations and subsequent interactions with oxygen. Visible light irradiation at ≥ 435
nm can excite the methylene blue electronic transition at 665 nm and initiate the
homogeneous self-degradation. Such side reactions can be minimized if another organic
compound is used, which does not have appreciable electronic transitions in the visible or
near-UV region. Under these conditions, the heterogeneous catalytic processes can be
detected more reliably. This strategy is explored by using acetone photodegradation as
described in the last section.
5.2. Methylene blue degradation by MoS2 and WS2 supported on SiO2
The relative absorption of methylene blue at 665 nm was plotted over time during
photodegradation with MoS2/SiO2, WS2/SiO2, and with the SiO2 control. Regression
analysis by Excel yielded best-fit lines for each. Each data set in Figure 23 contains three
data points at Time = 0 hours. However, not all of the symbols are visible due to overlap.
These three data points effectively weight each fit line toward the y-intercept of 1.0.
.
49
Methylene Blue Degradation with a Long-Pass 400 nm Filter
by Silica Deposited Catalyst
1.2
1.1
1
A/A0
Control
0.9
0.8
Molybdenum
Sulfide
0.7
Tungsten Sulfide
0.6
0.5
0
2
4
6
8
10
Time (Hrs)
Figure 23. Methylene blue degradation by MoS2/SiO2, WS2/SiO2, and with the SiO2 control. Relative
UV-visible absorption at 665 nm plotted vs. time.
From the initial examination of the plot, both MoS2/SiO2 and WS2/SiO2 seemed to
catalyze the methylene blue photodegradation compared to the SiO2 control. From the
raw data, it appears that the silica-supported MoS2 might be somewhat more active than
the WS2.
The standard deviations of the slopes were again calculated using the Linest function
in Excel to give the margin of error at the 90% Confidence Level. These are listed in
Table 9. The R-squared value of 0.6649 for the control plot suggests that there is much
scatter in the data, so the error in the slope may be too large to statistically determine that
the different slopes are not a result of random error. The error range for the control is
nearly twice that of either of the other two trials.
50
Table 9. Analysis of the rates of the methylene blue degradation by MoS2/SiO2 and WS2/SiO2. Slopes of
the linear fits from Figure 23 are listed along with the error limits at the 90% confidence level.
WS2/SiO2
Control (SiO2)
MoS2/SiO2
R2 Value
0.9386
0.8873
0.6649
-2
-1
Slope (x10 hr )
-2.7 ± 0.4
-2.3 ± 0.45
-1.9 ± 0.8
Upper Limit (x10-2 hr-2.3
-1.9
-1.1
Lower Limit (x10-2 hr-3.2
-2.8
-2.7
Due to that, at the 90% Confidence Level, the slopes of the WS2 and MoS2 runs already
fall within the error limits of the control. Therefore, without additional data processing,
we cannot say that either MoS2/SiO2 or WS2/SiO2 catalyzes the degradation of methylene
blue at the 90% confidence level.
5.3. Methylene blue degradation by MoS2 and WS2 supported on SiO2: Accounting
For Light Scattering by SiO2
Data from MoS2/SiO2, WS2/SiO2, and the SiO2 control were adjusted to account for
the light scattering by the solids. In a separate experiment, light scattering was
determined for MoS2/SiO2, WS2/SiO2, and SiO2 aqueous suspensions (see Table 7).
Absorbance at 665 nm was plotted against the absorbance at 420 nm for all three systems
(Figure 24). 665 nm is where methylene blue normally absorbs. 420 nm is where
methylene blue does not absorb and the only light losses here are due to scattering plus
any absorption by the deposited sulfide.
Table 10. Functions relating the scatter at 665 nm and 420 nm for MoS2/SiO2, WS2/SiO2, and SiO2.
Solid
MoS2/SiO2
WS2/SiO2
SiO2
Function
y = 0.5518x + 0.0224 y = 0.5178x + 0.0208 y = 1.0113x + 0.1963
R2 Value
0.9977
0.9684
0.967
51
Scattering Proportionality for Silica Deposited Tungsten Sulfide
665 nm scatter
0.2
0.19
0.18
0.17
0.16
0.15
0.14
0.25
0.26
0.27
0.28
0.29
0.3
0.31
0.32
0.33
0.34
0.35
420 nm scatter
Scattering Proportionality for Silica Deposited Molybdenum Sulfide
665 nm scatter
0.208
0.206
0.204
0.202
0.2
0.198
0.196
0.316
0.318
0.32
0.322
0.324
0.326
0.328
0.33
0.332
0.334
420 nm scatter
665 nm scatter
Scattering Proportionality for Silica
0.261
0.26
0.259
0.258
0.257
0.256
0.255
0.254
0.253
0.445
0.446
0.447
0.448
0.449
0.45
0.451
0.452
420 nm scatter
Figure 24. Linear correlation between light intensity losses at 665 nm and at 420 nm for aqueous
suspensions of three solids: (A) MoS2/SiO2, (B) WS2/SiO2, (C) SiO2.
52
Regression analysis by Excel calculated best-fit lines for each. The solids, their
equations, and the R-squared values along with the corresponding trials are listed in
Table 10.
The linear correlations yield different slopes for each of the three solids, which is to
be expected. Light scattering in a particular direction depends on the wavelength, the
particle size, and the refraction index, which also depends on the absorption spectrum,
according to Raleigh Theory.[17] As the MoS2 or WS2 nanoparticles attach to the SiO2
nanoparticles, the resulting particle size is larger and there is a change in the absorption
spectrum as well as the refraction index.
The found spectral correlations were then applied to the absorption measurements for
the reaction mixtures containing methylene blue and MoS2/SiO2, or WS2/SiO2, or the
SiO2 control. Scatter contribution at 665 nm was calculated for each spectral
measurement for each system. The calculated scatter contributions were then subtracted
from their respective measurements yielding a more correct absorbance of methylene
blue. Figure 25 displays these corrected kinetic data. Regression analysis was used to
determine best-fit lines. The slopes all increase in magnitude because now the A0 and A
values have been reduced by subtracting the scattering contribution from each. While
dividing by a smaller number, any change in the numerator will make a larger change in
the quotient.
53
Methylene Blue Degradation with a Long-Pass 400 nm Filter
by Silica Deposited Catalyst Adjusted for Light Scattering
1.2
1.1
A/A0
1
0.9
Control
0.8
Molybdenum
Sulfide
0.7
0.6
Tungsten Sulfide
0.5
0
2
4
6
8
10
Time (Hrs)
Figure 25. Methylene blue degradation by MoS2/SiO2, WS2/SiO2, and with the SiO2 control adjusted for
light scattering. Relative UV-visible absorption at 665 nm plotted vs time.
Using WS2 as a catalyst now appears to degrade methylene blue even faster than using
MoS2, which still degrades methylene blue faster than the silica control.
Statistical analysis was performed with the excel function Linest to find the standard
deviation for each slope and calculate the margin of error at the 90% Confidence Level.
Table 11. Analysis of the rates of the methylene blue degradation by MoS2/SiO2 and WS2/SiO2. Slopes of
the linear fits from Figure 25 are listed along with the error limits at the 90% confidence level.
MoS2/SiO2
WS2/SiO2
Control (SiO2)
2
R Value
0.9725
0.9767
0.6394
Slope (x10-2 hr-1)
-4.4 ± 0.46
-5.14 ± 0.435
-2.7 ± 1.24
Upper Limit (x10-2 hr-3.9
-4.70
-1.4
-2
Lower Limit (x10 hr
-4.8
-5.57
-3.9
There is an overlap in the error limits for MoS2/SiO2 and the error limits for the control at
the 90% Confidence Level. However, the slopes in the plots for the use of MoS2/SiO2
54
and WS2/SiO2 as catalysts are significantly more negative than the slope of the control
plot at the 90% Confidence Level. Further, the slope of the WS2/SiO2 plot is
significantly more negative than that of the MoS2/SiO2 plot. These facts indicate that
both sulfides successfully catalyzed the degradation of methylene blue under light ≥ 400
nm, but WS2/SiO2 worked significantly better as a catalyst than MoS2/SiO2.
5.4. Methylene blue degradation by MoS2 and WS2 supported on TiO2-Anatase
The relative absorption of methylene blue at 665 nm was plotted over time during
photodegradation with MoS2/TiO2, WS2/TiO2, and with the TiO2 control.
Degradation of Methylene Blue with a Long-Pass 400 nm
Filter by Titania Deposited Catalyst
1.2
1.1
Molybdenum
Sulfide
Control
A/A0
1
0.9
Tungsten Sulfide
0.8
0.7
0.6
0.5
0
1
2
3
4
5
Time (Hrs)
Figure 26. Relative UV-visible absorption at 665 nm plotted over time using titania and titania deposited
MoS2 and WS2 as catalysts.
55
Regression analysis by Excel yielded best-fit lines for each. The control run, TiO2, had
scattering that changed throughout the experiment, and this changing contribution to
absorption is reflected in the R-squared value of only .4796. Although WS2/TiO2 has a
visually more negative slope, it is still useful to analyze the error in each slope using the
Linest function to determine if there is any statistical difference between the slopes of
each run at the 90% Confidence Level.
Table 12. Analysis of the rates of the methylene blue degradation by MoS2/TiO2 and WS2/TiO2. Slopes of
the linear fits from Figure 26 are listed along with the error limits at the 90% confidence level.
MoS2/TiO2
WS2/TiO2
Control (TiO2)
2
R Value
0.8176
0.9976
0.4796
Slope (x10-2 hr-1)
-0.950 ± 0.3303
-2.4 ± 0.1
-0.959 ± 0.735
Upper Limit (x10-2 hr-0.620
-2.3
-0.224
Lower Limit (x10-2 hr-1.281
-2.5
-1.694
The MoS2/TiO2 and TiO2 slopes are statistically identical because the slope of each falls
into the error limits for the slope of the other. The slope of WS2/TiO2 is significantly
more negative than that of the others, and its error limits do not overlap the error limits of
the other two trials at this confidence level. So, this data indicates a significant increase
in the rate of methylene blue degradation at the 90% Confidence Level when WS2/TiO2 is
used in place of TiO2.
5.5. Methylene blue degradation by MoS2 and WS2 supported on TiO2-Anatase:
Accounting For Light Scattering by TiO2.
Data from MoS2/TiO2, WS2/TiO2, and the TiO2 control were adjusted to account for
the light scattering by the solids. In a separate experiment, light scattering was
determined for MoS2/TiO2, WS2/TiO2, and TiO2 aqueous suspensions (see Table 7).
56
Absorbance at 665 nm was plotted against the absorbance at 375 nm for all three systems
(Figure 27). 665 nm is where methylene blue normally absorbs. 375 nm is where
methylene blue does not absorb and the only light losses here are due to scattering plus
any absorption by the deposited sulfide. The best-fit function for each trial was
determined by regression analysis in Excel.
Scatter Correlation of Titania Deposited Catalysts
0.9
0.8
Scatter at 665 nm
0.7
Titania
0.6
0.5
0.4
Molybdenum
Sulfide
0.3
0.2
Tungsten Sulfide
0.1
0
0
0.2
0.4
0.6
0.8
1
Scatter at 375 nm
.
Figure 27. Linear correlation between light intensity losses due to scattering at 665 nm and at 375 nm for
aqueous suspensions of three solids: (A) MoS2/TiO2, (B) WS2/TiO2, (C) TiO2.
The line function and the R-squared value for each is displayed in Table 13 for easy
reference. The disparity between the scatter caused by the same mass of these different
solids reflects the need to adjust the titania-deposited sulfide trials for scattering. This
difference is not simply due the higher molecular weight of the deposited sulfide since
the mass ratio of sulfide to titania is only 1:100. It could be due to a difference in the
57
ability for the solid to disperse in a given amount of time spent stirring, particle size, or
refraction index as pointed out in the case of SiO2 scattering.
Table 13. Functions relating the scatter at 665 nm and 375 nm for MoS2/TiO2, WS2/TiO2, and TiO2.
WS2/TiO2
TiO2
Solid
MoS2/TiO2
Function
y = 0.8056x + 0.0021 y = 0.712x + 0.008
y = 0.8549x + 0.0481
R2 Value
0.9982
0.9981
0.9971
The functions for MoS2/TiO2, WS2/TiO2, and TiO2 scattering from Table 13 were
then applied to the absorption measurements at 375 nm for MoS2/TiO2, WS2/TiO2, and
TiO2 respectively. This resulted in a calculated scatter contribution for each
measurement of 665 nm absorption in these three degradation runs. The calculated
scatter contributions were then subtracted from their respective measurements. For each
trial, the last measurement of original absorption was used as the first data point and A0
to avoid using an average of growing scatter. The MoS2/TiO2 and WS2/TiO2 have more
negative slopes than the control, TiO2. The best fit line for the control has a higher Rsquared value now that the scatter has been subtracted. This data at 665 nm was plotted
in Excel as Figure 28.
58
Methylene Blue Degradation with a Long-Pass 400 nm Filter by
Titania Deposited Catalyst Adjusted for Light Scattering
1.2
1.1
A/A0
1
Control
0.9
0.8
Tungsten Sulfide
Molybdenum
Sulfide
0.7
0.6
0.5
0
1
2
3
4
5
Time (Hrs)
Figure 28. Methylene blue degradation by MoS2/TiO2, WS2/TiO2, and with the TiO2 control. Relative
UV-visible absorption at 665 nm has been corrected for the light scattering.
The MoS2/TiO2 and WS2/TiO2 have more negative slopes than the control, TiO2. The
most noticeable change is that the control data fits the function much better with an Rsquared value of 0.8682 compared to an R-squared value of 0.4796 before the scatter was
subtracted. Each slope has increased with the subtraction of the scatter from the A0 of
each trial, and so has the standard deviation of the slopes as calculated by the Linest
function in Excel.
Table 14. Error Analysis of the rates of the methylene blue degradation by MoS2/TiO2 and WS2/TiO2.
Slopes of the linear fits from Figure 28 are listed along with the error limits at the 90% confidence level.
MoS2/TiO2
WS2/TiO2
Control (TiO2)
R2 Value
0.9976
0.9900
0.8682
Slope (x10-2 hr-1)
-4.8 ± 0.22
-5.80 ± 0.554
-3.0 ± 1.1
Upper Limit (x10-2 hr-4.5
-5.25
-1.9
Lower Limit (x10-2 hr-5.0
-6.36
-4.1
59
As Table 14 shows, the error limits at the 90% Confidence Level do not overlap, which
means that these slopes are statistically different and the increase in the magnitude from
one slope to the next is significant. So, to the 90% Confidence Level, both MoS2/TiO2
and WS2/TiO2 significantly increased the rate of degradation of methylene blue compared
to the use of the TiO2 control. They succeeded in sensitizing the titania to visible light
for use as a catalyst. The increase in the magnitude of the slope for use of WS2/TiO2
compared to MoS2/TiO2 was also significant at this confidence level, meaning that
WS2/TiO2 exhibited more photocatalytic activity that MoS2/TiO2.
A comparison between WS2/SiO2 and WS2/TiO2 can be made (Table 12 and Table
14). Titania-supported WS2 exhibited 13% higher photocatalytic efficiency than its
silica-supported counterpart. The effect is probably due to the semiconducting nature of
TiO2, which allows for charge separation at the TiO2-WS2 interface. With MoS2, the
comparison is not so definitive. At a first glance, MoS2/TiO2 indeed exhibits a steeper
slope of the kinetic data than MoS2/SiO2, implying a higher photocatalytic activity.
However, given the scatter in the MoS2/SiO2 photocatalysis data, the slope of the
MoS2/TiO2 plot fits within the error limits of the slope for the MoS2/SiO2 plot at the 90%
Confidence Level. So, the slope using MoS2/TiO2 was not significantly more negative
than the slope using MoS2/SiO2 as a catalyst.
5.6. Acetone Degradation By Unsupported MoS2 and WS2
Unsupported MoS2 and WS2 catalyze photodegradation of acetone upon irradiation
with visible light of ≥ 400 nm wavelength. The absorption peak of acetone at 265 nm
progressively decreases (Figure 29).
60
Degradation of Acetone by Molybdenum Sulfide and
Light ≥ 400 nm Over Time with Baseline Subtracted
8.00E-02
Absorbance (arb. units)
7.00E-02
6.00E-02
5.00E-02
0 Minutes
62 Minutes
133 Minutes
204 Minutes
273 Minutes
390 Minutes
4.00E-02
3.00E-02
2.00E-02
1.00E-02
0.00E+00
200
-1.00E-02
220
240
260
280
300
320
Wavelength (nm)
Figure 29. The 265 nm absorption peak in the baseline corrected spectrum of acetone decreases as the
acetone/MoS2 mixture is irradiated by light ≥ 400 nm.
Since the unsupported sulfides do not disperse in the aqueous solution, there is no light
scattering contribution to subtract from the spectra. Due to the UV source variations in
the spectrometer, the acetone spectra are often slightly offset with respect to each other.
The offset can be removed by subtracting the baseline as measured at 320 nm.
The spectrum at 390 minutes may or may not indicate non-isosbestic behavior. It is
difficult to conclude if an isosbestic point is present in this short wavelength region
because random variations in the source make this region unreliable.
61
For MoS2, WS2, and the catalyst-free control solution, the baseline corrected
absorbance measurements were plotted in Excel and the best-fit lines were found by
regression analysis (Figure 30).
Degradation of Acetone with a Long-Pass 400 nm Filter by
Metal Sulfide Catalysts
1.1
1.05
A/A0
1
Control
0.95
Tungsten Sulfide
0.9
0.85
Molybdenum
Sulfide
0.8
0
1
2
3
4
5
6
7
Time (Hrs)
Figure 30. Baseline corrected, relative UV-visible absorption at 265 nm plotted over time using
unsupported MoS2 and WS2 as the catalysts.
From the plot, it is noticeable that the two metal sulfides successfully catalyze the
acetone degradation compared to the control mixture because they exhibit more negative
slopes than the control.
Using the Excel Linest function to calculate the standard deviation in the slope, the error limits at the
90% Confidence Level were calculated (
Table 15).
62
Table 15. Analysis of the rates of the acetone degradation by unsupported MoS2 and WS2. Slopes of the
linear fits from Figure 30 are listed along with the error limits at the 90% confidence level.
MoS2
WS2
Control (No Catalyst)
R2 Value
0.9839
0.8514
0.7924
-2
-1
Slope (x10 hr )
-2.69 ± 0.28 -1.84 ± 0.634
-1.2 ± 0.6
Upper Limit (x10-2 hr-2.41
-1.21
-0.6
Lower Limit (x10-2 hr-2.97
-2.48
-1.8
From these it can be seen that although error limits overlap, the slopes of the MoS2
and WS2 runs are significantly more negative than that of the control at the 90%
Confidence Level. The slope of MoS2 is also significantly more negative than that of
WS2. So, it would be correct to say at the 90% Confidence Level that the use of MoS2 as
a catalyst significantly increased the degradation rate of acetone with visible light
compared to the use of WS2 as a catalyst, and that both sulfides catalyzed the degradation
of acetone relative to the use of visible light alone.
5.7. Conclusions
1. Pure MoS2 and WS2 nanoparticles are very hydrophobic, which presented a challenge
in conducting the photocatalytic measurements in aqueous solutions. Inability to disperse
in water likely lowers the sulfide-water interface area and lowers the overall efficiency of
interfacial charge transfer. As a result, any side reactions and variations of experimental
conditions can mask the photocatalytic reaction kinetics.
2. Tests of the degradation of methylene blue by the unsupported sulfides under visible
light (λ ≥ 435 nm) demonstrated the photocatalytic activity of WS2 but not that of MoS2,
likely due to sufficient photo-bleaching rate of methylene blue under intense light. In
experiments with acetone, which is less susceptible to photo-bleaching, photocatalytic
activity was observed with MoS2 and, to a lesser degree, with WS2.
63
3. Supporting the sulfide nanoparticles on SiO2 or TiO2 allowed dispersing them in the
aqueous medium and increased the overall rates of photocatalytic degradation of
methylene blue. Accounting for light scattering by solids was essential here.
4. When silica, an insulator, is used as a support, WS2/SiO2 exhibits higher photocatalytic
activity than MoS2/SiO2. Both succeeded in catalyzing the degradation of methylene
blue compared to the SiO2 control at the 90% Confidence Level.
5. When titania, a semiconductor, is used as a support, WS2/TiO2 exhibits higher
photocatalytic activity than MoS2/TiO2, and both of these composite catalysts were
successful at increasing the degradation rate of methylene blue compared to TiO2 at the
90% Confidence Level. The titania-supported WS2 was significantly more photoactive
than the silica-supported WS2 at the 90% Confidence Level. This was probably due to
the semiconducting nature of TiO2, which allows for charge separation at the TiO2sulfide interface. Even though the titania-supported MoS2 appeared at first glance to be
more photoactive than the silica-supported MoS2, the increase in the degradation rate of
methylene blue was not significant at the 90% Confidence Level.
64
References
1. Fujishima, A.; Rao, T. N.; Tryk, D. A. Titanium Dioxide Photocatalysis.
J. Photochem. Photobiol. C 2000, 1, 1-21.
2. Mills, A.; Lee, S.-K. Semiconductor Photocatalysis, in Advanced Oxidation
Processes for Water and Wastewater Treatment, Parsons, S. (Ed.), IWA
Publishing, 2004.
3. Oppenländer, T. Photochemical Purification of Water and Air, Wiley-VCH, 2003.
4. Thurston, T.R.; Wilcoxon J.P. J. Phys. Chem. B 1999, 103, 11-17.
5. Wilcoxon, J. P.; Samara, G. A. Phys. Rev. B 1995, 51, 7299-7302.
6. Kelley, David F. Molecular and Supramolecular Photochem. 2003, 10, 173-209.
7. Wilcoxon, J. P. J. Phys. Chem. B 2000, 104, 7334- 7343.
8. Tributsch, H.; Bennett, J. C. J. Electroanal. Chem. 1977, 81, 97-111.
9. Ho, W. et al. Langmuir 2004, 20, 5865-5869.
10. Zong, X. et al. J. Am. Chem. Soc. 2008, 130, 7176–7177.
11. Duphil, D. et al. Nanotechnology 2004, 15, 828-832.
12. Strohmeier, W. Zeitschrift fuer Naturforschung 1962, 17b, 566.
13. King, R. B. Applied Spectroscopy 1969, 23, 536.
14. Brown, D. A.; Hughes, F. J. J. Chem. Soc. A 1968, 7, 1519-23.
15. Price, J. T.; Sorensen, T. S. Can. J. Chem. 1968, 46, 515-22.
16. Atkins, P. Physical Chemistry, 5th Ed.; W. H. Freeman and Co.: New York, 1994;
pp 993-999.
17. Atkins, P. Physical Chemistry, 5th Ed.; W. H. Freeman and Co.: New York, 1994;
pp 800-801.